4 ms·
I teach medical students and residents this method, and use it in my practice. There are some important caveats that make it more complicated than it looks at f
by phren0logy 12y ago
I teach medical students and residents this method, and use it in my practice. There are some important caveats that make it more complicated than it looks at first glance. The primary one is that the NNT is tied to a very specific outcome; for example: How many patients do I need to treat with a statin to lower cholesterol? That's a lower number than the number I need to treat go get the outcome I actually care about: How many patients do I need to treat with a statin to prevent one heart attack or stroke? And even then, the answer I actually get is: How many patients do I need to treat with a statin to prevent a heart attack or stroke in a one year period?
So, I would respectfully submit that it's not so simple as the article makes it sound, but it adds much-needed empiricism to prior standards for clinical decision-making.
- zarify 12y agoIf they made it sound more complicated then it'd be harder to tie a 'one weird trick' style headline to it ;P
- danielbarla 12y agoI also wonder if the situation isn't further complicated when you consider accuracy of diagnoses. A snippet from wikipedia: "It is estimated that between 10-15 percent of physician diagnoses are erroneous" [1], which is apparently attributed to overconfidence. This would probably have a larger effect in conjunction with treatments having a higher NNH (number needed to harm). In any case, I suspect that the push for (a more) evidence-based medical science will only have positive effects, and thanks for your information! [1] Berner, E.S., and M.L. Graber. "Overconfidence as a cause of diagnostic error in medicine." American Journal of Medicine 121 (2008): S2-S23.